Atrial fibrillation (AF) is the most common arrhythmia encountered in clinical practice, affecting over 2.2 million Americans and growing. Paroxysmal atrial fibrillation (PAF) is a self-terminating form of the cardiac arrhythmia that comprises between 25% and 60% of AF cases. PAF may be a single episode or clusters of abnormal rhythms with variable duration. PAF is often asymptomatic, and in those patients with symptoms, it has been estimated that for every symptomatic episode there are 10 asymptomatic ones. The self-terminating nature of PAF leads to its frequent under-diagnosis and consequent lack of use of established stroke prevention therapies. The proposed informatics screening tool ties clinical data collected in the home directly to the clinician; is completely unobtrusive and does not require surgical implantation. This system could identify even brief (5 minute) episodes of atrial fibrillation and prevent significant morbidity and mortality by identifying individuals wo would benefit from therapy, such as anticoagulation that significantly reduces stroke risk. The innovation of this project is a practical solution for providing such screening. A prototype device will be developed and evaluated in phase 1. The device will enable a significant reduction in morbidity and mortality in the US population.